: Chemotherapy is effective in prolonging patient survival and decreasing tumor recurrence for an increasing number of cancers. Almost all chemotherapy agents are associated with toxicity, which is the prime factor limiting their use. Current information on toxicity from different types of chemotherapy usually comes from randomized controlled trials. However, there have been few population-based assessments of toxicities associated with cancer chemotherapy in the community. Our studies on the external and internal validity of Medicare claims for chemotherapy suggest that Medicare claims data provide valid information on cancer chemotherapy. We have also demonstrated the feasibility of using Medicare claims data to assess population-based hospitalizations for serious toxicities associated with breast cancer chemotherapy. We propose to use the updated National Cancer Institute's SEER (Surveillance, Epidemiology, and End Results)-Medicare linked data to address in-depth issues on the postmarketing surveillance of short- and longterm toxicities associated with chemotherapy for several major cancers that are often treated with chemotherapy, including breast, ovarian, colorectal, and lung cancers. Our specific aims are: 1) to determine the rate of hospitalization for short-term toxicity associated with chemotherapy administration;2) to determine the mortality associated with hospitalization for short-term toxicity from chemotherapy;3) to determine the incidence of long-term or late-stage toxicity associated with chemotherapy use (e.g., cardiac dysfunction, bone marrow failure, second malignancies, and cognitive impairment);4) to determine how hospitalization, mortality and long-term toxicity vary with type of cancer, and type and amount of chemotherapy;and 5) to examine how hospitalization, mortality and long-term toxicity vary by age, gender, ethnicity, co-morbidity, socioeconomic factors, geographical area, hospital and physician characteristics. These analyses will be conducted in a large population-based cohort of over 300,000 patients diagnosed with the above cancers at age 65 or older from 1992 to 2002 in thirteen SEER areas, over 30% of whom receive chemotherapy. Currently, there is no systematic approach for evaluating long-term toxicity of the marketed drugs. The proposed innovative use of the nationwide, population-based computerized Medicare claims data offer a promising and less expensive way of obtaining information on postmarketing drug toxicity in community settings. This database is especially unique for studying cancer chemotherapy and its toxicity because chemotherapy is among the few drugs that are covered by Medicare for past several decades.